Twenty-seven years across operators, suppliers and towers

Mandate

Make a network built for fast growth secure, cheaper to power and demonstrably resilient, beginning with the removal of a supplier that had come to be treated as high-risk.

Outcomes

• Replaced a high-risk supplier's radios on 4,100 sites and moved the packet core to two other suppliers, for €214m against a €196m plan.
• Cut network electricity from 448 GWh (2022) to 362 GWh (2025) while traffic rose 71%; energy per gigabyte fell 53%.
• Introduced quarterly live failover between core sites after the July 2024 outage; eight tests completed, the last five with no customer-visible interruption.
• Moved 151,000 of 168,000 copper customers to fibre or fixed wireless under a per-sector capacity cap.
• Grew an open-interface radio pilot to 240 sites; tender prices for the replacement programme came in about 18% below our cost model.

Decisions in context

I joined Corvalta when its network had grown faster than its records. The company had built 9,400 sites in eight years, with two radio suppliers, a core assembled in stages and documentation that lived largely in engineers' heads. Within three months the board asked me for a plan to remove a supplier that had come to be treated as high-risk, and the energy crisis of 2022 then more than doubled our electricity bill. My work since has been to answer both together: we replaced the core and the radios on 4,100 sites, switched off 3G two and a half years early and more than halved energy per gigabyte. After the July 2024 voice outage I changed how resilience is tested and reported to the committee. I also led the engineering side of moving 168,000 copper customers to fibre or fixed wireless. I am not a director of Corvalta or of its holding company.

Reason for the move

Current appointment; completing the 5G standalone core migration and the third-core-site decision in 2027.

Mandate

Cut the energy and diesel costs of an ageing tower estate and engineer the towers to host more tenants and heavier 5G equipment.

Outcomes

• Replaced diesel generators with solar panels and lithium batteries on 1,450 off-grid and weak-grid towers; generator running hours fell 64%.
• Reassessed 3,200 towers for 5G antenna loads, strengthening 610, which supported a rise in tenants per tower from 1.62 to 1.91.
• Introduced battery autonomy classes based on what each site served, now the company's specification standard.
• Cut the energy and maintenance budget by 11% in real terms over four years.

Decisions in context

Torrevia had been carved out of a mobile operator two years before I arrived, and its commercial team was rightly focused on adding tenants to existing towers. I was hired to make the estate ready for them and cheaper to run. That meant a structural reassessment of towers for heavier 5G antennas, a programme to replace diesel generators at off-grid and weak-grid sites with solar panels and lithium batteries, and a battery standard set by what each site served rather than one figure for all. Studying energy systems part time at the Politecnico Albrenta gave me the vocabulary to argue with our power suppliers on their terms. I reported the energy programme to the board's sustainability committee each year. I left when Corvalta, one of our anchor tenants, asked me to run its network, a move I declared to Torrevia's chief executive, Enrico Malfante, as soon as the conversation began.

Reason for the move

Recruited by Corvalta Mobile, one of Torrevia's anchor tenants, to run its network; I declared the approach to Torrevia's chief executive when it began.

Mandate

Build the 4G network and negotiate, then run, a radio-sharing agreement for the interior of Portugal with a rival operator.

Outcomes

• Took 4G population coverage from launch to 97% by 2016.
• Negotiated a fifteen-year radio-sharing agreement with Marévia Móvel covering 1,900 sites, with exit terms agreed before cost-sharing terms.
• Decommissioned 640 duplicate sites, saving about €22m a year in operating costs.
• Presented the technical case and exit terms of the sharing agreement to Aldera's board in 2014.

Decisions in context

Aldera hired me from the supplier side to build its 4G network and to negotiate a radio-sharing agreement with Marévia Móvel for the interior of Portugal, where neither operator could justify a network of its own. The rollout kept to schedule. The sharing negotiation took eight months, most of them spent on governance, cost allocation and exit terms rather than on engineering. We decommissioned 640 duplicate sites and saved about €22 million a year in operating costs. In 2014 I presented the technical case and the exit terms to Aldera's board, my first appearance before a board. Working for an operator after nine years at a supplier showed me how much an operator's engineers know about their own network that never reaches a tender, and how much bargaining power an operator gives away when it writes a contract without them. I returned to Italy in 2017.

Reason for the move

Returned to Italy to lead engineering at a tower company.

Mandate

Run a Greek mobile operator's radio and transmission network under a seven-year managed-services contract, meeting its service levels at falling cost.

Outcomes

• Met the contract's network availability commitments in every year of my tenure.
• Absorbed four successive fee reductions totalling 21% while traffic grew, by renegotiating which work could wait.
• Planned island summer capacity two seasons ahead, cutting August congestion complaints by about half.
• Kept technician attrition below 9% a year through three years of frozen pay by promoting from within.

Decisions in context

Halvorsk ran the radio and transmission network of a Greek mobile operator under a seven-year managed-services contract, and I led the 340 people who did it, about a third of whom had transferred from the operator. Eighteen months after I arrived the financial crisis began, and the client reduced the fee every year while traffic grew. I renegotiated the service levels so that the cuts fell on work that could wait, such as cosmetic site maintenance, rather than on restoration times, and I planned summer capacity for the islands two seasons ahead. Keeping technicians through three years of frozen pay was the hardest part of the job, and we lost fewer than I feared because we promoted from within. I left wanting to be accountable for a network as its owner rather than as its contractor.

Reason for the move

Moved to the operator side to lead a radio network directly.

Mandate

Integrate new 3G radio and core equipment with operators' existing networks, then deliver rollout projects to time and acceptance.

Outcomes

• Integrated 3G equipment into two operators' networks with existing 2G systems.
• Delivered about 2,300 3G sites for two operators between 2005 and 2008.
• Introduced a site-readiness check before installation crews were dispatched, cutting repeat visits by about a third.
• Agreed acceptance tests with both operators that later became the Italian business's standard.

Decisions in context

I joined Halvorsk's Italian business as an integration engineer while operators were building their first 3G networks, and spent two years making new radio and core equipment work with customers' existing systems. From 2005 I managed rollout projects for two operators, with 45 engineers and a dozen installation subcontractors. The job taught me how a supplier allocates scarce engineers between customers, how acceptance tests are negotiated and why a contract that names no owner for the interfaces between systems will produce faults that nobody owns. I also learnt to refuse an operator's project manager when a site was not ready for installation, which made me unpopular for a week and saved weeks later. In 2008 Halvorsk asked me to take on its contract in Greece, my first role running a whole network.

Reason for the move

Asked to lead the company's managed-services contract in Greece.

Mandate

Plan new GSM sites, and later the first 3G sites, in Abruzzo and Molise, from coverage prediction to landlord agreement.

Outcomes

• Planned about 420 sites over three and a half years.
• Introduced drive-test verification within thirty days of every new site going live.
• Reduced the share of sites needing a second mast to cover their planned area from about 15% to 6%.

Decisions in context

My first job was planning radio sites for Nerantel's GSM network, and then its first 3G sites, in Abruzzo and Molise. I worked from prediction tools, drive tests and a great deal of time on hillsides with landlords and municipal officials. My first site above L'Aquila missed one of the two villages it was meant to cover, and it took eight months to add a second mast. I learnt to trust measurements over maps, and to plan for the permit, the access road and the power connection as carefully as for the coverage. I coordinated two drive-test teams and saw how much of a network's quality depends on field engineers who report what they actually see. I left because I wanted to understand how the equipment I was planning around worked inside, and the only place to learn that was at a supplier.

Reason for the move

Joined an equipment maker to understand how the equipment I was planning around actually worked.